| RATURI HIMANSHU | M, 2回目発表 | 数理情報学 | 池田 和司, | 安本 慶一, | 久保 孝富, | 日永田 智絵, | LI YUZHE |
|
Title: Towards Real-Time Tracking of Individuals in Crowded Environments
Abstract: Finding and tracking specific individuals in dense crowds—such as lost children, elderly persons, or patients—remains a critical challenge for public safety and rescue operations due to heavy occlusions and non-linear motion patterns. We propose a modular, multi-modal framework that integrates YOLOv8 for high-speed detection, MMPose (RTMPose-M) for structural keypoint analysis, and a Swin Transformer for robust visual re-identification. Unlike standard trackers that rely solely on visual bounding boxes, our system utilizes a Temporal Motion Transformer (TMT) to capture long-range movement dynamics and a Cross-Modal Fusion mechanism to dynamically prioritize reliable cues when appearance is degraded. Evaluation on the challenging MOT20 benchmark demonstrates that our approach achieves a MOTA of 69.6% and an IDF1 of 58.3%, effectively maintaining identity continuity even in extreme densities exceeding 100 pedestrians per frame. This work provides a scalable, real-time foundation for automated monitoring in complex, high-stakes public environments. Language of the presentation: English | |||||||
| ALNAJJAR MOHAMAD | M, 2回目発表 | ソーシャル・コンピューティング | 荒牧 英治, | 安本 慶一, | 若宮 翔子, | PENG SHAOWEN | |
|
title: *** Demographic-Specific Communication In Medical Text Using Large Language Models ***
abstract: *** Social media platforms such as Twitter and Facebook offer powerful channels for governments and organizations to share important public information that can support people in their daily lives. These platforms are especially valuable during times of crisis or danger—such as public health emergencies—when rapid and clear communication is essential. However, such posts often fail to reach or engage all segments of society. Differences in age, educational background, and other demographic factors influence how individuals perceive and interact with these messages. Some users may skip the posts entirely, while others may read them but struggle to understand the content. In this research, we aim to address this communication gap by identifying the underlying reasons why certain groups do not engage with medical content on social media. We then leverage large language models (LLMs) to adapt and rewrite these posts, improving their clarity, accessibility, and relevance. Our goal is to enhance the directivity of social media communication—ensuring that critical information reaches and resonates with diverse audiences across society. *** language of the presentation: *** English *** | |||||||
| YAO KOUAME JEAN FLORENTIN | M, 2回目発表 | ユビキタスコンピューティングシステム | 安本 慶一, | Sakriani Sakti, | 諏訪 博彦, | 佐々木 航, | 松井 智一 |
|
title: Dynamic Dark Pattern Detection in Mobile Applications
abstract: Dark patterns are deceptive interface design strategies that manipulate users into making decisions against their own interests. However, most existing detection approaches focus on static interface elements, while many contemporary dark patterns emerge over time through complex interaction flows. Therefore, this research investigates dynamic dark pattern detection by analyzing temporal interaction sequences using goal-oriented GUI agents and semantic flow modeling. language of the presentation: English | |||||||
| SCHINDLER CAROLIN | D, 中間発表 | ユビキタスコンピューティングシステム | 安本 慶一, | Sakriani Sakti, | 諏訪 博彦, | 松田 裕貴, | 佐々木 航 |
|
title: Towards Natural Multimodal Interaction with Argumentative Dialogue Systems
abstract: Argumentation is an essential element of human communication and reasoning. Whenever information is incomplete or inconsistent, we start to argue. This can take place internally as well as externally with other people or conversational agents. Recent works increase the flexibility of argumentative dialogue systems with respect to the possible topics of the argumentation. Still, the naturalness of the interaction with the dialogue system shows room for improvement. This research has identified three research objectives for improving the naturalness and multimodality of argumentative dialogue systems that can be integrated into existing computing systems: The automatic generation of argumentation structures, the presentation of the embodied argumentative dialogue system to the user, and the processing of freely formulated user utterances including unknown arguments. Within this presentation, special focus lies on ARCADE, an augmented reality platform for ubiquitous and natural interaction, and the estimation of subjectively perceived persuasion effectiveness, a research that is currently ongoing. language of the presentation: English | |||||||